Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
1998
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 1998

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: What is the fastest machine for Mathematica?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg13606] Re: What is the fastest machine for Mathematica?
  • From: dek at socrates.cgl.ucsf.edu (David Konerding)
  • Date: Fri, 7 Aug 1998 03:08:09 -0400
  • Organization: UCSF Computer Graphics Lab
  • References: <6q3qkj$f19@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

In article <6q3qkj$f19 at smc.vnet.net>, Christian Keysers wrote:
>We are currently running mathematica on a Silicon Graphics Ingigo 2
>machine with a 200Mhz processor and 256Mb Ram. The calculations involve
>fitting of gaussians to data set, and take two long at present (2h30min
>for each data set requiring 16 * 8 fits). Does anybody know what
>machines may bring the best speed for that kind of computations? The
>benchmark sites available in the net only contain data for relatively
>slow machines (Dec Alphas, Mac, Pentiums). We were thinking of getting
>an Origin 2000, but even these machines only get really fast with
>multiple processors (not supported by mathematica). So if anybody has a
>really fast machine, let me know!

DEC Alpha -- slow machine!?  I suspect the benchmarks (which I've never
looked at) are a bit out of date.  Newer DEC Alphas- such as the EV6
600MHz, which should be available pretty soon- should handily exceed
the speed of your Indigo2/200MHz. However, an O2000 would be no slouch
either.  Compare the specFP measures for the Alpha 600MHz and the
R10000/250MHz-- that's a decent measure for your task, which would be
primarily floating point and memory IO bound.

However, I suspect that if you really really wanted to speed execution
of your data fitting, some optimized C code could outperform
Mathematica.  There will always be some overhead associated with
running Mathematica code; I don't know how significant this will be, it
really depends on the routines you are using.

Dave
-- 
--------------------------------------------------------------------------------
Email: dek at cgl.ucsf.edu    David Konerding     WWW:
http://picasso.ucsf.edu/~dek
--------------------------------------------------------------------------------
Snail: Graduate Group in Biophysics
Medical Sciences 926, Box 0446
University of California
San Francisco, CA 94143


  • Prev by Date: RE: Re: Undocumented 3.0 Featu
  • Next by Date: Re: What is the fastest machine for Mathematica?
  • Previous by thread: Re: What is the fastest machine for Mathematica?
  • Next by thread: Re: What is the fastest machine for Mathematica?